INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

The maintenance of body potassium (K⁺) balance is critical to the normal function of all cells. Perturbations in K⁺ homeostasis disrupt normal cell growth and division, metabolism, volume and osmotic regulation, acid-base economy, and the excitability of nerve and contractile cells. The kidney is the principal arbiter of body K⁺ balance in mammals, adjusting K⁺ excretion to match large variations in dietary K⁺ intake. The late distal tubule and collecting duct have the dual ability to secrete and reabsorb K⁺ as needed to effect this balance (1, 2). In response to chronic dietary K⁺ deprivation, these segments, in particular the OMCD¹ actively reclaim filtered K⁺. Physiological, biochemical, and molecular biological studies have shown that this adaptation is principally attributable to increased expression and/or activity of an H⁺-K⁺-ATPase in the luminal membrane of these segments (2-9). A similar transport system(s) has been identified in the apical membrane of mammalian distal colon, where it, too, effects active K⁺ absorption (10, 11). Although active K⁺ absorption in the distal colon is enhanced during K⁺ depletion (11) and participates to a limited degree in restoring K⁺ balance, the identity of the specific K⁺-ATPase that is up-regulated remains controversial (12).

The H⁺-K⁺-ATPases constitute a subfamily of isozymes that belong to the X⁺-K⁺-ATPase multigene family, which also includes the Na⁺-K⁺-ATPase isoforms. The X⁺-K⁺-ATPases share common catalytic and ion transport mechanisms and an apparent requirement for heterodimeric (:) structure. The X⁺-K⁺-ATPase  subunits exhibit considerable (~65%) structural homology and contribute most of the functional properties of the holoenzymes, but they can be distinguished to a degree from one another on the basis of organ distributions and sensitivities to the inhibitors ouabain and Sch 28080 (13). To date, three distinct H⁺-K⁺-ATPase  subunits have been cloned from mammals, structurally characterized, and expressed in heterologous systems. The H⁺-K⁺-ATPase 1 subunit (HK1) was first cloned from and is principally expressed in stomach (14), where it plays a major role in gastric acid secretion. Messenger RNA encoding this gene was also identified in the renal collecting duct (3). The pharmacological signature of the HK1 protein is its high sensitivity to inhibition by Sch 28080 and its complete resistance to inhibition by ouabain (15, 16). The H⁺-K⁺-ATPase 2 cDNA was first cloned from rat distal colon (17), where it is abundantly expressed, and lower levels of HK2 mRNA were reported in proximal colon (17), uterus (17), and kidney (5-8). Expression of the HK2 subunit with the known rat X⁺-K⁺-ATPase  subunits (16) or toad bladder H⁺-K⁺-ATPase  subunit (18) in Xenopus laevis oocytes resulted in the appearance of active H⁺-K⁺ exchange that was virtually resistant to Sch 28080 and partially inhibited by ouabain. When HK2 was expressed without an exogenous  subunit in Sf9 cells, the resultant K⁺-ATPase activity was Sch 28080- and ouabain-resistant (19). A third H⁺-K⁺-ATPase  subunit cDNA, termed ATP1AL1 (or H⁺-K⁺-ATPase 4), was cloned from a human skin cDNA library (20), and transcripts encoding this gene product were also detected in human brain and kidney but not colon (20). Coexpression of the ATP1AL1 subunit and the rabbit gastric H⁺-K⁺-ATPase  subunit (HK_g) in Xenopus oocytes (21) or HEK 293 cells (22) resulted in the expression of functional H⁺-K⁺ pumps that were partially sensitive to both Sch 28080 and ouabain.

Recent studies by our laboratory and others have shown that chronic K⁺ deprivation enhances HK2, but not HK1 (3), gene expression in the OMCD (5-8) and proximal portion of the inner medullary collecting duct (6) of rats. In one of these studies (8), HK2 protein levels, but not mRNA levels, were enhanced in the outer medulla of K⁺-deprived rats, suggesting the potential operation of translational or post-translational control mechanisms. In contrast to kidney, chronic hypokalemia does not appear to alter HK2 mRNA (5, 8) or protein (8) abundance in rat distal colon. Moreover, recent work demonstrating disparate effects of adrenalectomy, dexamethasone treatment (5), and dietary Na⁺ depletion (8) on HK2 abundance in the rat outer medulla and distal colon indicated that cell type-specific regulatory mechanisms govern HK2 gene expression in these tissues.

Since both transcriptional and translational control mechanisms, as well as alternative mRNA splicing, can lead to regulated, tissue-specific gene expression, we hypothesized that these mechanisms might operate to confer structural and/or regulatory diversity to the HK2 subunit gene. Although the structural organization of the rat and human HK1 (23) and human ATP1AL1 (24) genes is known, that of the rat HK2 gene has not been described. We report here that distinct transcription initiation sites in the rat HK2 gene and alternative mRNA splicing, combined regulatory mechanisms not known to be utilized by other members of the X⁺-K⁺-ATPase  subunit family, direct the synthesis of two N-terminal HK2 variants that are expressed principally, if not exclusively, in the kidney and colon and that appear to respond coordinately in kidney to chronic K⁺ deprivation.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Animal Protocols-- Male Sprague-Dawley rats (180-220 g) were fed normal rat chow (150 mEq KCl/kg chow, TD88081, Harlan Teklad) or a nominally K⁺-free (TD88082, Harlan Teklad) diet for 2 weeks. This K⁺-restriction protocol reproducibly results in significant hypokalemia (3) and has been used in our previous studies (3, 6).

Oligonucleotide Primers-- PCR primers not included in specific kits were synthesized by Genosys, Inc. (The Woodlands, TX). The sequences of the various HK2 subunit primers are presented in Fig. 1A, and those of the HK_g subunit are given below.

5-RACE and Cloning of H⁺-K⁺-ATPase 2b cDNA-- The 5-RACE protocol was performed using the Marathon^TM cDNA Amplification Kit (CLONTECH, Palo Alto, CA), according to the manufacturer's instructions. First strand cDNAs were generated from 1 µg of rat kidney poly(A)⁺ RNA, using Moloney murine leukemia virus reverse transcriptase and a modified locking oligo(dT) primer containing two degenerate nucleotide positions at its 3 end provided with the kit. Second strand synthesis was accomplished with a mixture of Escherichia coli DNA polymerase I, RNase H, and E. coli DNA ligase. After creation of blunt ends with T4 DNA polymerase, the double-stranded cDNA was ligated to adapter primer 1 furnished with the kit, using T4 DNA ligase. The anchor-ligated cDNAs were then subjected to 5-RACE using a nested primer (adapter primer 2, supplied with the RACE kit) complementary to adapter primer 1, HK2-specific reverse primer (P1, Fig. 1A) complementary to nucleotides +344 to +325 of the published HK2 cDNA sequence (17), and the components of the Advantage^TM cDNA Amplification Kit (CLONTECH). PCR cycling conditions were as follows: 94 °C × 1 min, followed by 28 cycles of 94 °C × 30 s, 68 °C × 4 min, and a final step of 68 °C × 4 min. Ten µl of the amplified products were separated by electrophoresis in a 1% agarose gel and visualized by ethidium bromide staining and UV shadowing. The final amplicons were then subcloned into the plasmid vector pCR2.1^TM (Invitrogen) and sequenced on both strands by a cycle sequencing method.

Cloning of the Rat HK_g Subunit cDNA-- The encoding DNA of the rat HK_g subunit was PCR-amplified from rat stomach cDNA using primers flanking the coding region: sense 5- ATAAGCTTCAGCCCTGCAGGAGAAG-3 (nucleotides +16 to +32 of the published sequence (25)) and antisense 5- ATTCTAGATTACTTCTGTATTGTGAGC-3 (nucleotides +878 to +896 of the published sequence). HindIII and XbaI sites (underlined) were added to the 5 ends of the sense and antisense HK_g primers, respectively, to facilitate subcloning. The resultant amplicon was digested with HindIII and XbaI and subcloned into these sites of the mammalian expression vector pcDNA3.1+/Zeo (yielding the recombinant pcDNA3.1/HK_g-Zeo). The insert HK_g DNA was sequenced to verify its authenticity.

Primer Extension-- Antisense primers (Fig. 1A) specific for HK2a (P7) and HK2b (P8 and P9) were 5-end-labeled with [-³²P]ATP using T4 polynucleotide kinase. The primers were annealed to 10 µg of total RNA from distal colon at 58 °C for 20 min. After cooling at room temperature for 10 min, the primers were extended with avian myeloblastosis virus reverse transcriptase at 42 °C for 15 min in a reaction mixture containing 50 mM Tris-HCl, pH 8.3, at 42 °C, 50 mM KCl, 10 mM MgCl₂, 10 mM dithiothreitol, 1 mM each dNTP, 0.5 mM spermidine, and 2.8 mM sodium pyrophosphate. The reactions were stopped by the addition of gel loading dye, and the samples were heated at 90 °C for 10 min. The primer extension products were resolved by electrophoresis on 8% acrylamide, 7 M urea polyacrylamide gels in TBE buffer. The sizes of the primer extension products were established by comparison with a sequence ladder generated by cycle sequencing with the ³²P-labeled primer used for each extension reaction and the HK2 partial genomic DNA clone (see "Results") as template.

Analysis of Rat Genomic DNA-- The 5 end of the HK2 gene was analyzed using the Rat PromoterFinder^TM DNA Walking Kit (CLONTECH), which contains separate pools ("libraries") of uncloned, genomic DNA that have been predigested with EcoRV, ScaI, DraI, PvuII, or SspI and ligated to an oligonucleotide anchor (adapter primer 1). A nested PCR approach was employed. In the first round, aliquots of each "library" were amplified with adapter primer 1 and HK2 primer P1, using a program of 94 °C × 25 s, 72 °C × 4 min for 7 cycles, 94 °C × 25 s, 67 °C × 4 min for 32 cycles, and 67 °C × 4 min for 1 cycle. After analysis of an aliquot of the PCR products on a 1.2% agarose gel, the remaining PCR products were diluted 1:50 in sterile deionized H₂0 and subjected to a second round of PCR, using the nested adapter primer 2 and the nested HK2 primer P2 (Fig. 1A) in a program of 94 °C × 25 s, 72 °C × 4 min for 7 cycles, 94 °C × 25 s, 67 °C × 4 min for 20 cycles, and 67 °C × 4 min for 1 cycle. The amplified products were separated by electrophoresis in a 0.9% agarose gel, subcloned into pCR2.1^TM, and sequenced on both strands by a cycle sequencing method.

RNA Isolation and Northern Analysis-- Total RNA was extracted from selected tissues and renal parenchymal zones of normal and K⁺-deprived rats using RNAzol B (Tel-Test). The samples were quantitated by spectrophotometry at 260 nm. Isoform-specific cDNAs of roughly comparable length (Fig. 1A) were generated by PCR from the cloned HK2a and HK2b cDNAs, using primer pairs P3 + P4 and P5 + P6 (Fig. 1A) directed at the unique 5 exonic sequences of the HK2a and HK2b isoforms, respectively. Sequence analysis showed that these regions exhibited no significant homology to each other or to any sequence in the GenBank data base. A rat GAPDH cDNA (nucleotides 469-984, Ref. 26) was also generated by PCR. For Northern analysis, the GAPDH and HK2a- and HK2b-specific cDNAs were radiolabeled with ³²P by the random primer method according to the manufacturer's instructions (Prime-a-Gene, Promega, Madison, WI). Fifteen µg of total RNA per lane were separated by size on 1% agarose, 2% formaldehyde gels and blotted to nylon membranes (Hybond N, Amersham Corp.). After UV cross-linking, the blots were visualized under UV light, hybridized for 2 h at 68 °C in QuickHyb solution (Stratagene) with probes specific for HK2a, HK2b, or GAPDH (as an additional control for RNA quality and equality of sample loading and transfer), and washed to a final stringency of 0.1 × SSC, 0.1% (s/v) SDS at 60 °C. Autoradiographs of the blots were prepared at 70 °C. In several experiments (as indicated in the figure legends), the blots were sequentially hybridized with HK2a and HK2b DNA probes of comparable size and specific activity, followed by the GAPDH DNA probe, with the blots being stripped before proceeding to the next analysis. After each stripping, autoradiographs of the blots were prepared to verify removal of the probe.

In Vitro Transcription and Translation-- pcDNA3.1/HK2b-Neo and the HK2a encoding DNA subcloned into the vector pAGA2 (16) were transcribed and translated in the presence of [³⁵S]methionine with T7 RNA polymerase and the TNT-coupled reticulocyte lysate kit (Promega, Madison, WI). The synthesized proteins were separated by SDS-polyacrylamide gel electrophoresis and analyzed by fluorography.

Cell Culture and Transfection-- HEK 293 cells were grown in modified Eagle's medium containing 10% fetal bovine serum, 100 units/ml penicillin, 100 µg/ml streptomycin, 5 µg/ml insulin, 5 µg/ml transferrin, 5 ng/ml selenium, and 2 mM L-glutamine (complete medium). Subconfluent HEK 293 cells grown on 10-mm culture dishes were transfected with pcDNA3.1/Neo (as a vector control) or pcDNA3.1/HK2b-Neo with the Tfx-50 reagent (Boehringer Mannheim) to yield HEK-NEO and HEK-HK2b cell lines, respectively. In brief, 10 µg of plasmid DNA and 22 µl of Tfx-50 reagent were mixed with 6 ml of modified Eagle's medium. The mixture was added to the monolayers and incubated for 2 h at 37 °C in a 5% CO₂ incubator. Twelve ml of prewarmed complete medium was then overlaid onto the medium, and the cells were returned to the incubator. After 48 h, the medium was replaced with complete medium containing 600 µg/ml G418 (Life Technologies, Inc.). The G418-containing medium was replaced every 3 days until individual resistant colonies were isolated and established in culture as individual lines. All lines were maintained in G418 medium and frozen after one to three in vitro passages. HEK-HK2b clone 25 was used in the functional analysis detailed below. To test whether coexpression of the HK_g affected functional expression of HK2b, HEK-NEO and HEK-HK2b cell lines were stably transfected with pcDNA3.1/HK_g-Zeo and selected in complete medium containing 600 µg/ml G418 and 250 µg/ml Zeocin. Cells surviving selection were screened for HK2b and/or HK_g expression by Northern analysis with probes specific for each subunit. The doubly transfected cells were termed HEK-HK2b/HK_g, and clone 40 was selected for further functional analysis.

⁸⁶Rb⁺ Uptake-- Uptake of ⁸⁶Rb⁺, a K⁺ congener, was measured at 37 °C in transfected HEK 293 cells grown in 24-well plates according to a published protocol (22). Monolayers were rinsed five times and preincubated in uptake buffer (145 mM NaCl, 1 mM KCl, 10 mM glucose, 1.2 mM MgCl₂, 1 mM CaCl₂, 2 mM NaH₂PO₄, 32 mM HEPES, pH 7.4, and 200 µM bumetanide) at 37 °C for 20 min in the presence or absence of different concentrations of ouabain as indicated in the figure legends. External 1 mM K⁺ was used in these assays, because K⁺ competitively inhibits both Sch 28080 and ouabain binding to X⁺-K⁺-ATPase  subunits, and this concentration is within the narrow range of K_m values reported for K⁺ dependence of all known X⁺-K⁺-ATPase  subunits. Uptake was initiated by adding 0.2 ml of uptake buffer containing ~4 µCi/ml ⁸⁶Rb⁺. After 12 min, the reaction was stopped by six rapid washes with ice-cold stop buffer (100 mM MgCl₂, 10 mM Tris-HEPES, pH 7.4). Parametric studies indicated that this time point was in the linear range of uptake. The cells were solubilized in 2% SDS, 0.1 N NaOH, and the resulting extracts were measured for ⁸⁶Rb⁺ by Cerenkov radiation and for protein content by the BCA Protein Assay Reagent (Pierce). Triplicate or quadruplicate measurements were obtained in each uptake condition.

Data Analysis-- The intensities of bands on the Northern blot autoradiograms were measured by whole band densitometry software running on a SPARC Station IPC (Sun Microsystems, Mountain View, CA) equipped with an image analysis system (BioImage, Ann Arbor, MI). Predictions of membrane-spanning regions and their orientation were generated by the TMpred program (27) through the ISREC Bioinformatics Group server. Predictions of potential promoter regions were obtained with a neural networks algorithm (28) through the LBNL Human Genome Informatics Group server. Potential regulatory motifs in the HK2 gene were identified with Transcription Element Search Software from the Computational Biology and Informatics Laboratory server of the University of Pennsylvania School of Medicine, using the Transfac 3.1 data base. Quantitative data are presented as mean ± S.E. and were analyzed for significance by analysis of variance. Significance was assigned at p < 0.05.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

cDNA Cloning and Structural Analysis of a Truncated N-terminal Variant of the H⁺-K⁺-ATPase 2 Subunit-- The anchor-ligated cDNAs synthesized from rat kidney mRNA were subjected to 5-RACE using adapter primer 2 and HK2-specific primer P1 from exon 2 (Fig. 1A). Two distinct PCR products of ~400 and ~600 bp, subsequently shown to correspond to the 5 ends of HK2a and HK2b, respectively, were consistently amplified. These products were isolated, subcloned, and sequenced. A total of 16 RACE reactions for both amplicons was analyzed in this manner. The two RACE product subtypes differed in sequence at their 5 ends but were identical at their 3 ends, with common sequence beginning at the codon for Lys⁴ of the known HK2a sequence (Fig. 1A). HK2a was identical in sequence to the corresponding region of the HK2 cDNA reported by Crowson and Shull (17) but included an additional 72 bp at its 5 end, so that the total 5-UTR was 274 bp.

View larger version (50K): [in this window] [in a new window]   Fig. 1.   5 end of the rat H+-K+-ATPase 2 (HK2) gene. A, nucleotide sequence, potential elements, and PCR primer sequences. (In the discussion and Fig. 1, we have assumed that the 5-most exon in the H+-K+-ATPase 2 gene is exon 1. However, the possible existence of additional upstream exons has not been rigorously excluded.) The 5-flanking regions and intron 1 sequences are indicated by lowercase letters. The sequences of the HK2a and HK2b transcription units are indicated in uppercase black and white letters, respectively. The 5 splice donor and 3 splice acceptor sites are identified. Amino acids encoded by exons 1 and 2 are indicated below the appropriate codons. Nucleotides used to generate oligonucleotide primers for PCR and primer extension are italicized, and their orientation (sense, antisense) is indicated by arrows. Potential TATA and CACCC sequences are doubly underlined, and CCAAT sequences are indicated by open boxes. The transcription initiation sites defined by primer extension are designated by arrows. 5-Most ends of the 5-RACE sequences are indicated by a caret over the nucleotide. Sequences exhibiting homology to cis-elements or transcription factor binding sites (AP-2, AP-3, Sp1, GATA-1, NF-B, PEA-3, C/EBP, GR, IRF-1, NF-interleukin 6, and HNF-4) are underlined. Numbers to the right of the figure indicate nucleotide positions relative to the HK2a translation start site. Potential sites of protein kinase A and protein kinase C phosphorylation are designated by asterisks. B, schematic representation showing alternative splicing of HK2a and HK2b mRNAs. Exons are depicted as boxes, with the shaded area representing the sequence spliced from the HK2a mRNA and retained in the HK2b sequence. The translation initiation codons of the major open reading frames are indicated by AUG. The positions of the exon groups and splicing patterns were identified by a PCR cloning method as detailed under "Experimental Procedures." The putative first promoter precedes exon 1, which contains a start ATG and only 3 codons. The putative second promoter resides in intron 1.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Secondary structure analysis and predicted structural models of H+-K+-ATPase 2a (HK2a) and H+-K+-ATPase 2b (HK2b) isoforms. A, TMpred analysis (27) of the HK2a and HK2b amino acid sequences using a window of 17-33 amino acids for the length of the hydrophobic part of the transmembrane helix. B, schematic representation of the membrane topology of HK2a and HK2b. The truncated portion for HK2b is indicated by shading. Putative protein kinase A and protein kinase C phosphorylation sites are indicated by P.

View larger version (43K): [in this window] [in a new window]   Fig. 3.   In vitro transcription and translation of H+-K+-ATPase 2a (HK2a) and H+-K+-ATPase 2b (HK2b). Fluorograph of SDS-10% polyacrylamide gel of peptides generated, as described under "Experimental Procedures," by in vitro transcription and translation of HK2a and HK2b encoding DNAs in the presence of [35S]methionine. , indicates no DNA template. Molecular weights were determined by comparisons of their mobilities to known standards. Results are representative of two experiments.

Genomic Organization of the 5 End of the HK2 Gene and Mapping of the Transcription Start Sites-- Analysis of the HK2a and HK2b 5-RACE products suggested that both mRNAs are derived from a single gene by utilization of alternative splice sites at the 5 region. To determine the order of the exons and the intervening genomic sequences, we used nested reverse primers derived from the 5 region common to both variants and adapter-ligated rat genomic DNA libraries to PCR amplify a portion of the 5 region of the HK2 gene. PCR products of ~1.6- and 0.5-kb were consistently amplified from the DraI and PvuII libraries, respectively. These amplicons were subcloned into pCR2.1^TM and sequenced on both strands. Sequence analysis indicated that the HK2b transcript is the product of an alternative transcription initiation site located within the first intron (Fig. 1, A and B). The 5 end of the HK2b mRNA represents a 5 extension of exon 2 that is excised in the HK2a mRNA. In support of this construct, a consensus 5 splice donor site (5-GTGAGT-3) was identified at the exon 1/intron 1 boundary, and a consensus 3 acceptor site (CAG) (29) was found in the expected 5 region of exon 2 (Fig. 1A).

View larger version (49K): [in this window] [in a new window]   Fig. 4.   Primer extension analysis of 5 ends of H+-K+-ATPase 2 isoform mRNAs. Primer extension experiments were performed with 32P-labeled HK2 isoform-specific oligonucleotide primers P7 (HK2a) and P8 and P9 (HK2b, see Fig. 1A), respectively, and total RNA from rat distal colon as described under "Experimental Procedures." Representative autoradiographs for the HK2a and HK2b (primer P9 results) results are shown in A and B, respectively. Identical mapping results were obtained with the HK2b primers P8 and P9. Yeast tRNA served as a negative control (). Lanes 1-5 represent RNA samples obtained from 5 different rats. Lanes A, C, G, and T are sequencing reactions on the same gel using the same primer and the plasmid construct bearing the HK2 1.6-kb genomic fragment obtained from the DraI-digested rat genomic DNA library (see "Experimental Procedures"). The base corresponding to the major transcription start site for each isoform is labeled by an asterisk within the genomic DNA sequence shown to the right. The slightly slower mobility of the primer extension products in lanes 3-5 of A reflects a slight delay in loading of these samples. In both figures, short and long exposures of the film were used to allow optimal comparison of the sequencing ladders with the primer extension reactions (which were run on the same gel). All experiments were performed in triplicate.

Analysis of Potential Gene Control Elements-- The HK2 partial genomic clone contained sequences of ~380 and ~205 bp immediately 5 to the transcription start sites of the HK2a and HK2b transcription units, respectively. These sequences share no obvious homology, and they were examined for potential DNA elements that may contribute to transcriptional initiation and regulation. The HK2a 5-flanking region contains a TATA-like sequence (ATTTAA), a CACCC sequence (31), and a CCAAT motif (30) beginning 27, 127, and 133 bp, respectively, 5 to the transcription start site, which likely comprise the core promoter module (Fig. 1A). The 380 bp immediately preceding the transcription start sites contains several potential cis-elements that may serve as binding sites for transcription factors. These include 7 Sp 1 sites (32), 3 AP-2 sites (33), 2 GR sites (31), and single GATA-1 (34), C/EBP (35), PEA-3 (36), NF-B (37), and HNF-4 (38) motifs (Fig. 1A).

HK2 Isoform mRNAs Are Expressed in Colon and Kidney-- Northern blots of total RNA harvested from an array of tissues harvested from K⁺-replete rats were probed with ³²P-labeled DNA probes specific for each HK2 subtype (Figs. 5 and 6). Both isoforms were expressed prominently in the distal colon (Fig. 5) and very weakly in the proximal colon and normal kidney (Fig. 6). No transcripts were detected in skeletal muscle, heart, brain, stomach, spleen, liver, testis, or lung (Fig. 5), even with prolonged autoradiographic exposures. Failure to detect transcripts in these latter tissues also indicates that the HK2 isoform-specific probes did not cross-hybridize with the four known Na⁺-K⁺-ATPase  subunit isoforms (abundantly expressed in heart, brain, skeletal muscle, and/or testis (41, 42)) or the HK1 subunit (abundantly expressed in stomach (14)). Moreover, reprobing the blots with a ³²P-labeled DNA probe for GAPDH indicated comparable abundance and integrity of the blotted RNA samples (data not shown).

View larger version (65K): [in this window] [in a new window]   Fig. 5.   Tissue distribution of H+-K+-ATPase 2a (HK2a) and H+-K+-ATPase 2b (HK2b) mRNA in rat by high stringency Northern analysis. Total RNA (15 µg) from the indicated rat tissues was electrophoresed in 1% agarose-formaldehyde gels and transferred to nylon membranes as detailed under "Experimental Procedures." The filters were probed sequentially with 32P-labeled DNA probes specific for HK2a and HK2b (see Fig. 1 and "Experimental Procedures").

View larger version (38K): [in this window] [in a new window]   Fig. 6.   H+-K+-ATPase 2a (HK2a) and H+-K+-ATPase 2b (HK2b) subunit mRNA expression in colon and renal parenchymal zones of control (C) and potassium-restricted () rats. A, representative Northern blot of total RNA from cortex (CTX), outer medulla (OM), inner medulla (IM), proximal colon (PC), and distal colon (DC) isolated from control and K+-restricted rats (n = 4 animals for each group). The filters were probed sequentially with 32P-labeled DNA probes specific for HK2a, HK2b, and GAPDH as described under "Experimental Procedures." Autoradiographic exposure was overnight. B, histogram showing results of densitometric analysis of Northern blots. The ratio of the relative optical density of the HK2b and HK2a transcript bands in the K+-restricted rats is plotted.

Expression of HK2 Subunit Isoform mRNAs in Kidney and Colon of Control and K⁺-restricted Rats-- To determine the response of the HK2a and HK2b gene products to dietary K⁺ restriction in rat kidney and colon, Northern analysis with ³²P-labeled DNA probes specific for each HK2 subtype was performed on total RNA harvested from the proximal colon, distal colon, and renal cortex, outer medulla, and inner medulla of control and K⁺-restricted rats. In control rats, an abundant ~4.0-kb subunit transcript (HK2b  HK2a) was detected with both the HK2a- and HK2b-specific probes in distal colon (Fig. 6A). In addition, the HK2b-specific probe hybridized to a much less abundant ~6.0-kb transcript in distal colon (Fig. 6A). It is not known whether this larger transcript represents a processing intermediate or an mRNA with an alternate polyadenylation signal, but similar results were reported by Crowson and Shull (17), who used C-terminal coding and 3-UTR sequences as probes. With prolonged autoradiographic exposures (3 days), very low, comparable levels of HK2a subunit mRNA were detected in the proximal colon, cortex, and outer and inner medulla (data not shown). When these blots were reprobed with the HK2b-specific probe of roughly comparable size and specific activity, detectable ~4.0-kb transcripts were observed in the same structures after overnight exposure, suggesting that HK2b is expressed at higher levels, albeit still very low, than HK2a in normal kidney and colon.

Functional Expression of the H⁺-K⁺-ATPase 2b Subunit in HEK 293 Cells-- A dual selection strategy, using separate mammalian expression vectors containing the encoding DNAs for HK2b and HK_g together with the neomycin and Zeocin resistance genes, respectively, was employed to generate cell lines stably expressing the HK2b subunit, the HK_g subunit, or both subunits. HEK 293 cells were chosen as the recipient cells for the transfection experiments because they are easily transfected, do not express H⁺-K⁺-ATPase  or  subunit gene products, their endogenous Na⁺-K⁺-ATPase is highly sensitive to ouabain (22), and they permit analysis of H⁺-K⁺-ATPase biosynthesis and subunit assembly in mammalian cells at 37 °C (a factor that has been suggested to influence the fidelity of oligomerization and membrane insertion of the pump, Ref. 22).

View larger version (29K): [in this window] [in a new window]   Fig. 7.   Heterologous expression of the H+-K+-ATPase 2b subunit with or without the gastric H+-K+-ATPase  subunit (HKg) in HEK 293 cells. A, autoradiograph of representative Northern blots of total RNA harvested from HEK-NEO cells (NEO) and HEK-HK2b (2b) clone 25 cells. The blot was probed with a 32P-labeled DNA probe specific for HK2b. B, autoradiograph of Northern blot of total RNA harvested from HEK-NEO and HEK-HK2b/HKg (g) clone 40 cells. The latter cell line was generated by stable transfection of HEK-HK2b clone 25 cells with the DNA encoding HKg. The blot was probed with a 32P-labeled DNA probe specific for HKg. The minor, higher molecular weight bands presumably represent processing intermediates or differences in polyadenylation (provided by the bovine growth hormone poly(A) sequence included in the pcDNA3.1+/Zeo vector) of the HKg mRNA.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Effect of ouabain and Sch 28080 concentration on 86Rb+ uptake in HEK 293 cells stably expressing the H+-K+-ATPase 2b subunit with or without the gastric H+-K+-ATPase  subunit (HKg) subunit. A, 86Rb+ uptake in HEK-NEO, HEK-HKg, HEK-HK2b clone 25, or HEK-HK2b/HKg clone 40 cells was determined in the presence of 1 mM KCl, 200 µM bumetanide, and 1 µM or 1 mM ouabain. Data (means ± S.E.) are expressed as the percent of the control value (presented under "Results") measured in the absence of ouabain (n = 3). *, p < 0.05 compared with comparably treated HEK-NEO and HEK-HKg cells. The apparent difference between the HEK-HK2b clone 25 and the HEK-HK2b/HKg clone 40 cells did not achieve statistical significance. B, 86Rb+ uptake in HEK-NEO and HEK-HK2b/HKg clone 40 cells was measured as a function of the ouabain concentration in the presence of 1 mM KCl and 200 µM bumetanide. Data (means ± S.E.) are expressed as the percent of the control value measured in the absence of ouabain (n = 3). Where error bars are not seen, they are contained within the datum point. Control values were 4.9 ± 0.3 nmol/mg protein/min for HEK-NEO cells and 4.5 ± 0.5 nmol/mg protein/min for HEK-HK2b/HKg clone 40 cells. *, p < 0.05 compared with HEK-NEO cells. C, 86Rb+ uptake in HEK-HK2b cells and HEK-HK2b/HKg clone 25 was determined in the presence of 1 mM KCl, 200 µM bumetanide, 1 µM ouabain (to inhibit endogenous Na+-K+-ATPase activity), and vehicle (control) or the indicated concentrations of Sch 28080 (means ± S.E. n = 3).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Analysis of the regulation of active K⁺ reabsorption in the renal collecting duct and distal colon has been hampered by the lack of structural data concerning potential control mechanisms governing HK2 gene expression. In this study, we characterized two alternatively spliced products of the rat HK2 gene, HK2a and HK2b, that apparently arise from the use of alternative promoters and differ in the length of their N termini and their relative abundance in kidney and colon. Heterologous expression studies of the novel transcript in HEK 293 cells indicate that HK2b encodes a plasma membrane mechanism for K⁺ uptake that, like that of the full-length HK2 subunit (16, 18, 19), is relatively Sch 28080-resistant, intermediate in its sensitivity to ouabain, and operates more effectively when coexpressed with the HK_g subunit. The HK2b isoform represents the most abundantly expressed HK2 transcript in the rat kidney and distal colon and the principal H⁺-K⁺-ATPase transcript up-regulated in the renal medulla of K⁺-deprived rats. We also identified structural features that may govern transcriptional initiation and control as well as translational regulation of these isoforms. Our results suggest that both HK2 isoforms may contribute to K⁺ conservation during chronic hypokalemia, and they uncover a new degree of regulatory complexity for the X⁺-K⁺-ATPase  subunit gene family.

The first variant, HK2a, is the previously described (17) 1036-amino acid protein, which is encoded by a 4.0-kb mRNA transcribed from the 5-most putative promoter. Primer extension analysis places the major transcription initiation site 274 bp upstream of the translation initiation methionine. Exon 1 includes the 5-UTR and encodes the first three amino acids of the primary HK2a translation product. This structural theme is common to other members of the X⁺-K⁺-ATPase  subunit family. The analogous exon in ATP1AL1 also encodes 3 amino acids, whereas those for the HK1, and Na⁺-K⁺-ATPase 1 and 2 subunits genes encode 4 amino acids, and that for the Na⁺-K⁺-ATPase 3 subunit gene encodes only 2 amino acids. The 5-flanking region of HK2a contains common basal promoter elements. An AT-rich sequence that might serve as a TATA element begins 21 bp 5 to the transcription start site. This sequence is preceded by potential CCAAT (30) and CACCC (31) elements residing within the preferred context for such elements. In addition, sequence inspection of the 5-flanking regions revealed potential cis-acting DNA elements, including sites for AP-2, AP-3, GATA-1, HNF-4, C/EBP, GR, PEA-3, NF-B, and multiple Sp1 sites sequences that may participate in transcriptional regulation of this gene. Of these, Sp1 (43) and GATA DNA-binding proteins (44) have been shown to play important roles in transcriptional activation of the HK1 gene. Since we did not confirm the 5 end of the HK2 gene, other potential regulatory elements may reside upstream of the sequence we characterized.

The second variant, HK2b, has not been previously recognized. This 929-amino acid protein is also encoded by an ~4.0-kb mRNA that is transcribed from an internal putative promoter residing in intron 1. Given the near-identical size of the major HK2a and HK2b mRNA transcripts, Northern analysis with probes directed to sites distal to the alternative splice site in exon 2 would be unable to distinguish between the two isoforms. Our DNA sequence data (Fig. 1A) combined with the in vitro transcription and translation results (Fig. 3) and functional expression data (Fig. 7, B and C) indicate that the HK2b isoform encodes a protein with the requisite features of an X⁺-K⁺-ATPase  subunit. Primer extension and 5-RACE identified a putative site for HK2b transcription initiation, but given the context of the surrounding nucleotides, additional transcription start sites may be located upstream of the site we identified (that is closer to the TATA, CACCC, and reverse complement CCAAT sequences found in the 5-flanking region of the HK2b transcription unit). Other potential cis-elements, including multiple AP-2 and Sp-1 sites, as well as consensus NF-interleukin 6, IRF-1, AP-4, GR, and GATA-1 sequences, were identified in this region. Conclusive evidence for the functional activity of the HK2a and HK2b promoter elements will require formal testing with promoter-reporter gene constructs.

A notable feature of the HK2b mRNA is the complex 5-UTR containing multiple, partially overlapping AUG triplets in ORFs upstream (uORF) of the translation initiation site of the major ORF. Recent analyses have shown that such uORFs are present in <10% of vertebrate mRNAs (45) and that in some instances they inhibit translational initiation at the major ORF. For example uORFs in the 5-UTR of the retinoic acid receptor 2 and transforming growth factor 3 mRNAs dramatically inhibited CAP-dependent translation in vitro (46, 47). Moreover, studies of the retinoic acid receptor 2 mRNA in transgenic mice indicate a role for uORFs in tissue-specific and developmentally regulated gene expression (48). The relatively low level of ⁸⁶Rb⁺ uptake activity attributable to the HK2b pump in the HEK-HK2b and HEK-HK2b/HK_g cell lines (Fig. 8A) despite abundant mRNA expression (Fig. 7A) might reflect this regulatory constraint. Alternatively, preference of HK2b for an X⁺-K⁺-ATPase  subunit other than the endogenous Na⁺-K⁺-ATPase 1 subunit expressed in HEK 293 cells might limit expression of transport activity. The fact that coexpression of the HK_g subunit supported higher rates of ⁸⁶Rb⁺ uptake activity and was required for survival in 1 µM ouabain supports this latter hypothesis. Since studies of the full-length HK2 subunit, expressed by cRNA injection in Xenopus oocytes, indicated that the rat Na⁺-K⁺-ATPase 1 and HK_g subunit support comparable rates of K⁺ uptake (16), it remains to be determined whether the two HK2 isoforms differ in their promiscuity for X⁺-K⁺-ATPase  subunits.

The functional and regulatory significance of the N-terminal truncation of HK2b remains to be explored in further detail. The N terminus is the most variable structural region among the X⁺-K⁺-ATPase  subunits. Conceivably the decision to code for the N-terminal 108 amino acids present in HK2a could dictate isoform-specific differences in membrane targeting or cytoskeletal association in polarized epithelia, regulation by protein kinase C or protein kinase A phosphorylation (since these sites are present in HK2a but not HK2b), ion transport kinetics, or inhibitor sensitivities. There is precedent for alternative promoters to direct the coding of protein variants that are targeted to different intracellular locales. The two variants of leukemia inhibitory factor, which exist as diffusible and extracellular matrix-associated isoforms, represent such an occurrence (49). The possibility for functional and pharmacological differences in the HK2 isoforms is particularly intriguing since N-terminal deletion mutants of the closely related Na⁺-K⁺-ATPase exhibited altered K⁺ deocclusion kinetics compared with wild-type pumps (50), and since the N-terminal truncation of HK2b impinges on the H1-H2 domains, which have been implicated in ouabain and Sch 28080 binding to other X⁺-K⁺-ATPase  subunits (reviewed in Ref. 13). However, like the full-length HK2 subunit (16, 18, 19), HK2b is insensitive to high concentrations of Sch 28080. Moreover, the approximate IC₅₀ for ouabain inhibition (400-800 µM in the presence of external 1 mM K⁺) of the HK2b/HK_g pump reported here is comparable to values reported for the full-length HK2 subunit expressed in heterologous systems. Codina et al. (16) reported an IC₅₀ of 400-600 µM in the presence of external 1 mM K⁺ for HK2 pumps expressed in Xenopus oocytes, and Cougnon et al. (18) reported K_i values for ouabain of ~70 and ~970 µM in the presence of external 0.2 and 5 mM K⁺, respectively, for HK2 pumps expressed in HEK 293 cells. Clearly, heterologous expression and detailed functional analysis of the two isoforms in a common host cell will be needed to distinguish subtle differences.

These considerations take on added meaning when viewed in the context of recent functional studies in kidney and colon. In vitro studies have identified at least three different K⁺-ATPase activities that are distinguished by their kinetic and pharmacological properties in rat kidney (51, 52). One activity (type I) is K⁺-, but not Na⁺-dependent, ouabain-resistant, Sch 28080-sensitive, and expressed in collecting ducts. A second activity (type II) is K⁺-, but not Na⁺-dependent, Sch 28080- and ouabain-sensitive, and expressed basally in proximal tubules and the thick ascending limbs (52). This activity is virtually abolished during chronic K⁺ depletion. A third activity (type III) is activated by either Na⁺ or K⁺, exhibits higher sensitivities to ouabain and to Sch 28080 than type II, and a lower sensitivity to Sch 28080 than type I. This activity is not expressed basally but is specifically up-regulated in cortical collecting ducts and OMCDs with chronic hypokalemia (52). Similarly, both ouabain-sensitive and insensitive K⁺-ATPase activities have been identified in the apical membranes of colonocytes from the distal colon (10), yet only HK2 mRNA (5, 8) and protein (8) have been identified in these cells. These collective data have led us to postulate that a yet-to-be discovered K⁺-ATPase isoform may be operative in the renal collecting duct and colon (8, 19, 51). It is possible that functional differences in the HK2 protein variants may account for these puzzling data.

In addition to the generation of protein isoforms differing at the N terminus, the use of alternative promoters in the HK2 gene would be expected to afford considerable versatility in controlling its expression. Alternative promoter usage in other genes has been shown to allow for expression of isoforms exhibiting differences in the degree and timing of transcription initiation, mRNA turnover, translational efficiency, tissue specificity, and responses to signal transduction pathways (53). The HK2 gene appears to be the first example of a P-type ATPase to employ alternative promoters and mRNA splicing to generate structural and regulatory diversity. This mechanism, then, adds to the known complexity of X⁺-K⁺-ATPase regulation, which includes controls on transcription, translational efficiency, subunit assembly, and various post-translational modifications. It may also provide an explanation for the well documented differential expression of the HK2 gene in kidney and distal colon under various experimental conditions. For example, Jaisser and co-workers (5) showed that chronic K⁺ deprivation did not alter, adrenalectomy reduced, and dexamethasone supplementation of adrenalectomized rats restored steady-state HK2 mRNA levels in distal colon. In contrast, chronic K⁺ deprivation enhanced HK2 mRNA expression in the OMCD, whereas adrenalectomy did not alter HK2 gene expression. We (6) and others (7) have shown similar effects of K⁺ deprivation on HK2 mRNA levels in the OMCD. The probes used in all these studies would be expected to hybridize to both HK2 variants. Similarly, Sangan and colleagues (8) showed that chronic dietary Na⁺ depletion (presumed to promote secondary hyperaldosteronism), but not chronic K⁺ depletion, enhanced HK2 mRNA and protein levels in distal colon. Conversely, chronic K⁺ depletion promoted HK2 protein but not mRNA expression in outer medulla, whereas Na⁺ depletion did not affect renal expression of this gene product. Fortuitously, the antibody (termed M-1) used in this and an earlier (19) study was raised against a fusion protein produced from the first 109 amino acids of the HK2a sequence. Thus, this antibody would be expected to be specific for HK2a, and it would not detect HK2b. M-1 immunoreactivity was identified in the apical membranes of principal cells of the K⁺-deprived OMCD (8) and of surface cells in rat distal colon (19). Since the consensus of in situ hybridization studies (5, 6) with probes common to the two HK2 variants indicated that HK2 mRNA is primarily expressed in OMCD intercalated cells, it is reasonable to hypothesize that the two isoforms are expressed in different cell types of the rat OMCD during K⁺ depletion. The sequence information presented here should facilitate future studies to define the molecular mechanisms controlling the differential and cell type-specific expression of these isoforms.

Finally, although it has been hypothesized that HK2 and ATP1AL1 represent species variants of the same protein, the novel structural organization and regulatory mechanisms for HK2 transcription described here add to the growing list of differences that suggest that these proteins represent distinct protein isoforms. These distinguishing features include the contrasting tissue distributions (20), differences in pharmacological profile (16, 18, 19, 21, 22), and greater degree of sequence divergence when compared with the interspecies differences of the other human and rat X⁺-K⁺-ATPase  subunit isoforms. As additional structure-function and structure-regulation correlations for these genes are identified, their evolutionary relationship should come into clearer focus.